A collection of model organisms of sandbagging, that is, deliberate underperformance on a task. Code is at https://github.com/sandbaggers/sandbaggers.
Lennart Finke
lennart-finke
AI & ML interests
AI Safety
Recent Activity
updated a dataset 2 days ago
lennart-finke/realvul5 updated a model 2 days ago
lennart-finke/Strong-TAR updated a collection 9 days ago
sandbaggingOrganizations
models 8
lennart-finke/Strong-TAR
35B • Updated • 28
lennart-finke/SD
36B • Updated • 13
lennart-finke/Reinforcement-Learning
36B • Updated • 16
lennart-finke/Expert-Iteration
36B • Updated • 12
lennart-finke/SDF
36B • Updated • 11
lennart-finke/Password-locked
36B • Updated • 15
lennart-finke/KTO-SFT
36B • Updated • 17
lennart-finke/Classifier
Updated